This R03 proposal seeks two years of funding to conduct secondary analyses on an important longitudinal dataset examining predictors of high-risk drinking and driving in a select sample of DWI offenders (N=1396). Secondary analyses provide a fruitful and cost-effective means of testing and refining theory and can augment existing knowledge using recently developed and innovative statistical techniques. Prior work examining alcoholic subtypes and classification of high-risk drinkers has emphasized the role of personality factors and driving-related attitudes as efficient proximal predictors. [unreadable] [unreadable] However, this body of knowledge has neglected examining the etiologic role of cognitive processes and psychiatric symptoms as they contribute to vulnerability. An accumulation of empirical evidence based on prospective alcohol-related studies highlights that cognitive and psychiatric factors play a prominent generative role in the early stages of drinking and contribute to heavy and more problematic abuse. However, more detailed studies are required that clarify whether specific risk factors moderate or mediate high-risk drinking and driving. The present application extends previous empirical findings and theoretical knowledge in several ways: (1) by including a wider set of demographic, cognitive, behavioral, attitudinal, and psychological (including diagnostic) predictors than previously considered; (2) by using a high-risk sample of convicted DWI offenders, with adequate representation of women and minorities, which provides a unique opportunity to learn more about etiology and consequences of problematic alcohol use in an offender population; (3) by examining 5-year histories of substance abuse treatment and its relation to the progression of alcohol and drug dependence symptoms; and (4) by implementing state-of-the-art multivariate statistical analysis methods including confirmatory factor analysis and structural equation modeling (SEM) to examine theoretical models of high-risk drinking. Participants were convicted drunk drivers referred to the Lovelace Comprehensive Screening Program from the municipal court system in Bernalilto County, New Mexico, contacted and interviewed five years after referral for screening. Data collection protocols included structured in-person interviews coupled with research diagnostic criteria for alcohol and drug, and other psychiatric disorders. This study has the potential to elucidate important risk mechanisms that contribute to high-risk drinking and driving and presence of psychiatric symptoms in the five years following a first DWI conviction, as well as elucidate a theoretically consistent foundation on which to construct clinically useful screening methods, develop valid prevention approaches, and design efficient treatment approaches. [unreadable] [unreadable]